GAB is pleased to publish this Guest Post by Maya Forstater, well-known analyst on business and sustainable development, on a topic of continuing concern to scholars and activists working on corruption and development matters.
Are unreliable guesstimates and made-up statistics mildly irritating, indispensably powerful or potentially dangerous in the public debates on corruption? The topic comes up so often on the Global Anti-Corruption Blog that it has been given its own own three-letter acronym: WAGs (or Wild Ass Guesses).
Those at the sharp end of advocacy maintain, with some justification, that in the battle for attention, an arrestingly big number makes all the difference. But as Rick has argued, overinflated figures can also cause harm.
Something similar happens on the related topic of tax and illicit flows. One example of this is the widespread belief that ‘developing countries lose three times more to the tax avoidance by multinational companies than they receive in aid’. This much quoted WAG gives the impression of huge potential gains for the poorest countries, but is based on a chain of misunderstandings . In practice the magnitudes of revenues at stake are likely to be several times smaller than aid for the countries where that comparison matters.
Similarly, broad estimates of illicit flows or the scale of the black economy (“trillions”) are often presented in ways that suggest that the sums to be gained from tackling corporate tax avoidance are larger than any serious analysis supports.
I have written about these big numbers previously in a paper published by the Centre for Global Development here (or here for the short version).
But what harm do such numbers do, compared to their power at getting people talking about the issues? Is it really worth pointing out misunderstandings and myths in pursuit of a more rigorous and careful approach to evidence? (Or as I have been asked‘ Do you ever wonder how much you help the tax abusers?’)
I see four key dangers from inflated perceptions of the numbers:
First, as Rick argues bad numbers debase political debate and undermine confidence in public institutions. In the UK for example the revenue authority HMRC recently reached a tax settlement with Google, which became the subject of furious speculation. A common headline was that Google had been given a 3% tax rate. Despite HMRCs explanation about why this calculation bore no relation to how tax law works, the perception stuck, contributing to a political storm which generated plenty of heat but little light. Similarly, Malawi has recently been used by international campaigns as an example of a country where multinational companies are ‘paying little or no corporate tax’, with the suggestion that the tax behavior of foreign investors is both the cause and potential solution for major shortfalls in funding for public services in Malawi. The danger of such exaggerated claims is that they may distract attention from priorities higher up on Malawi’s agenda (If you think this is an undue worry read this article in the Nyasa Times).
A second potential harm is the impact on”tax morale.” The honesty with which people conduct their own tax affairs is influenced by whether they perceive the system to be fair, and what they think others are doing. An exaggerated perception that big companies are paying little or no tax risks could discourage voluntary compliance and make it harder to collect tax across the board.
A third danger is the impact on the investment environment. Companies say that what matters most in taxation is not low rates, but stability and predictability. Unstable and erratically applied tax rules and incentives are the worst of all worlds – potentially deterring investment without reliably raising revenue. Zambia’s polarized political debate and shifting mining tax regime is one example. WAGs such as the much-repeated but never explained statement that ‘Zambia loses USD2 billion every year due to tax evasion and avoidance” added fuel to the fire of distrust between the Government and industry. This number, which has no published basis and is out of scale with the underlying economics of Zambian copper, continues to be enthusiastically cited internationally as a proof point for the huge scale of potential gains from clamping down on multinational tax avoidance.
A fourth harm is the impact on the institutions themselves. Rough estimates can be a good starting point for shining a light on issues. But such numbers can easily be made to appear larger than they are (such as by presenting estimates which relate mainly to large emerging economies as relating to ‘the poorest countries’, or by setting numbers calculated for a broad region such as Africa against more concentrated spending needs such as healthcare in Ebola effected countries ). Hyping the number weakens processes of organizational learning and integrity. The paradox of all this is that these big numbers are often being used to make the case for greater transparency, with the hope that it informs and empowers citizens to hold governments and the powerful to account. But for this to work there needs to be a robust chain of links between numbers and information, real analysis and understanding and sustained citizen engagement on complex issues.
Rick’s suggestion of ‘health warning label’ on unsupported numbers didn’t get much response back in January. I think what is needed is not so much a negative label, but positive brand credibility secured by organizations taking care that their presentation of the numbers does not mislead. This does not mean giving up on using rough estimates. But it does mean recognizing that they can easily slip into presentations which make them appear much bigger than they are
I certainly agree that a huge amount of integrity and support is lost when members of the anti corruption community purposefully use misleading statistics — such as lumping in estimates for the largest developing economies as if we were only speaking about the poorest countries.
I wonder however how much of that inflation in the name of advocacy is going on and how much of it is simply that a lot of these numbers are extremely difficult to measure accurately. I’d be curious what level of confidence you think we need to have in WAGs before it is productive rather than counterproductive to use them?
One other note: the impact on tax moral would occur from publicizing accurate numbers if a lot of people were avoiding taxes. Do you believe that even true numbers should not be reported in order to positively influence behaviors? (This wouldn’t be unheard of: political campaigns attempt to affect voting moral by telling people there will be record turnout even when they believe turnout is likely to be extremely low). That could actually be expanded into the moral of corruption generally — does making people aware of high levels of corruption increase corrupt behavior?
Thanks for your response and good questions.
There are of course difficulties with measurement, but I don’t think it is uncertainty which causes these misinterpretations to proliferate. Rather I think it is the strength of the narrative into which the numbers are fitted – i.e. that tackling multinational taxation will release “huge sums of money for development’’. There are available estimates (c $100- 200 billion for developing countries) which can be used cautiously, but to fit them into this narrative the presentation ends up being fudged in one of three ways:
– Just say ‘this is a huge amount’
– Compare estimates relating to all developing and emerging economies and public spending/ mortality or aid concentrated on lower income countries
– Fudge the distinction between gross illicit flow estimates and revenue numbers
I don’t think that these exaggerations in presentation are necessarily done on purpose, but they are widespread. What i see more often is that people have an expectation,that the numbers are large in relation to development needs so they tend to misinterpret the sources to fit in with that expectation. So for example one report may develop a careful estimate which legitimately includes large emerging economies under ‘developing countries’ (using the UN categorisation). But then someone else cites this estimate and assumes that ‘developing’ countries means something closer to ‘LDCs’ and illustrates it with photos and cases etc…which reflect this interpretation. The next person therefore interprets it as ‘the poorest countries’ and adds a comparison to aid volumes etc…
On tax morale – No I don’t think this is a reason not to publicise knowledge about tax avoidance (or corruption) with true numbers.
Pingback: The Panama Filing Cabinets | Hiya Maya
Pingback: Power in Numbers: How Data Shapes the Tax Policy Arena | fairskat blog
If you want to see further cases of corruption in Zambia look at the following blog:
It basically details how ZCCM-IH, a partly owned government mining company, has been overrun by corruption and nepotism recently. The company’s financial reports for 2016 noted a turnover of 199 million Kwacha and an operating loss of 846 million Kwacha – approximately US$84 million. This all indicates that major kickbacks and illicit payments have been made to government and high-ranking company officials.
The blog Enough Corruption is run by a team of independent investigative journalists who have also covered the Airbus SE case.
Pingback: How Reliable Are Global Quantitative Corruption Statistics? A New U4 Report Suggests the Need for Caution | GAB | The Global Anticorruption Blog